Researchers at the University of Massachusetts Amherst have discovered a "Zombie Card" attack that allows expired Visa contactless cards to be used for in-store purchases by manipulating the expiration date read by POS terminals over NFC, without compromising the card's cryptographic protections. The attack requires only physical access to the card and demonstrates a vulnerability in how contactless payment systems validate card expiration data during transactions.
Corero Network Security has integrated AI-Augmented Cloud-Assist into SmartWall ONE, adding cloud-based AI analysis and threat intelligence to its automated DDoS protection platform. The enhancement enables faster identification of emerging attack patterns and rapid generation of protective policies that can be deployed manually or automatically within seconds. This development reflects the industry's need for AI-driven defenses to match the sophistication of AI-powered attack campaigns.
AWS has introduced a method for maintaining user authorization context through AI agents, enabling infrastructure and downstream services to enforce access controls rather than depending solely on the agent itself. This approach addresses a critical security gap where agents without knowledge of the requesting user could potentially return unauthorized information, even when manipulated. Customers using Amazon Bedrock AgentCore can now build AI agents that safely access various data sources including DynamoDB tables, document repositories, and SaaS platforms while maintaining proper access restrictions.
Autonomous AI agents are now accessing sensitive enterprise data, writing code, and executing workflows once reserved for trusted employees, fundamentally shifting the attack surface and enterprise trust model beyond traditional cybersecurity controls. Mohammad Arif, Head of Information Security at Guild Group, warns that most organisations are treating AI security as a future problem despite widespread adoption already underway, and advocates for embedding AI governance into procurement, data protection, identity management, and incident response as a board-level priority. Over the next 12 months, enterprises should expect AI-assisted social engineering, sensitive data leakage into unapproved tools, insecure AI integrations with excessive permissions, and third-party AI supply-chain risks to materialise across their expanded attack surface.
At Black Hat this month, a group of security researchers took a $9,000 robot dog, plugged Google's AI into its brain, and jailbroke it by telling it - with a completely straight face - that it was a Pokemon. What followed involved a wall, a blue ice chest, and anyone in the room wearing white shoes. Oh, and did we mention you can buy a flamethrower attachment? Meanwhile, in Salzburg, 280 gold statuettes of Mozart have vanished from the streets. This has happened to the same artist before. Organised crime, or a publicity stunt? Jenny has thoughts - and some parallels for the world of cybersecurity. All this and more in episode 481 of the "Smashing Security" podcast with cybersecurity expert and keynote speaker Graham Cluley, and special guest Jenny Radcliffe.
A new AI platform called Kriminal is being offered without content filters or safety guardrails, despite official policies against illicit use, enabling bad actors to access tools for social engineering, offensive cybercrime operations, and reconnaissance scanning. The service's accessibility via cryptocurrency payments raises concerns about its potential abuse for coordinated cyberattacks and other malicious activities in the threat landscape.
Enterprises face a new insider threat vector as agentic AI systems gain adoption, requiring organizations to implement monitoring strategies for autonomous agents operating within their infrastructure. The concern has been heightened following recent security incidents, prompting security leaders to reassess how AI agents with system access could be exploited or compromised to facilitate attacks from within.
OpenAI has temporarily halted reinforcement learning training for its latest models to strengthen its internal safety defenses and expand monitoring capabilities, citing growing risks as AI systems become more advanced. The two-week pause follows concerns about preventing incidents similar to a previous Hugging Face-related event. The company emphasized that developing and testing increasingly capable models carries escalating risks that necessitate enhanced protective measures.
Microsoft is named a visionary leader in the 2026 Frost Radar for Cloud Workload Protection Platforms, recognized for unified runtime security with Microsoft Defender for Cloud. The post Microsoft named a Leader in the Frost Radar™: Cloud Workload Protection Platforms, 2026 appeared first on Microsoft Security Blog.
Wiz has released its Penetration Test Findings capability to general availability, enabling organizations to move beyond traditional one-time pen tests by integrating findings into continuous exposure management workflows. The unified platform merges penetration test results with real-time cloud context to provide ongoing visibility into vulnerabilities and exposures across cloud environments.
Rapid7 has partnered with Licencias OnLine to expand distribution of its AI-powered cybersecurity operations platform across Latin America, addressing the region's growing security challenges from cloud adoption and digital transformation. The partnership will provide technical training, partner enablement, and go-to-market support to help organizations reduce attack surface blind spots and strengthen their security operations maturity.
Flock’s surveillance cameras have already sparked outrage. WIRED reconstructed its next-generation AI system, already in use by some police, to confirm it goes much further than tracking license plates.
UT San Antonio delayed the start of fall classes by three days following a cyber incident detected over the weekend, with the semester now beginning August 24 to allow time for system restoration. The university's immediate response involved taking systems and services offline as a precaution, temporarily disrupting connectivity and email, though investigation found no evidence of data access or exfiltration. Technology teams working with expert partners continue restoring services and strengthening security safeguards across the campus network.
A 27-year-old data analyst from Charlotte, North Carolina, faced prison time after attempting to extort $2.5 million from his employer upon learning his contract would not be renewed. Rather than pursuing legitimate job search options, Cameron Curry chose extortion as his response to the employment setback. The case highlights how employment disputes can escalate into serious criminal activity with severe legal consequences.